English(EN) From Geometry to Behavior: How Instruction Exposure Unlocks Latent Rhetorical Directions in LLMs

指令暴露解锁大型语言模型中的类比向量

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-31 04:00

研究人员在大型语言模型中识别出一个代表潜在修辞策略的“类比向量”。该向量在模型的残差流中以几何方式可访问，可用于引导模型的解释风格。研究发现，虽然类比方向存在于所有基础模型中，但其行为可访问性随训练过程中指令暴露量的增加而扩展，这挑战了传统的基础模型与指令微调模型之间的二分法。 AI

影响揭示了指令暴露如何塑造大型语言模型的修辞能力，可能指导未来的模型训练和可解释性方法。

排序理由在arXiv上发表的研究论文，详细介绍了一种新的大型语言模型可解释性技术。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 在arXiv上发表的研究论文，详细介绍了一种新的大型语言模型可解释性技术。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: 10 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

arXiv cs.AI TIER_1 English(EN) · Shengjie Xu · 2026-10-31 04:00

从几何到行为：指令暴露如何解锁LLM中潜在的修辞方向

arXiv:2610.37685v1 Announce Type: cross Abstract: Language models do not merely memorize facts; they absorb rhetorical strategies from pretraining data, including the human tendency to reason by analogy. We ask: is this absorbed capability geometrically accessible before any inst…